RAND Corporation

Report Part Title: Unifying Tasking, Collection, Processing, Exploitation, and

Dissemination (TCPED) Across the U.S. Intelligence Community
Report Title: Perspectives and Opportunities in Intelligence for U.S. Leaders
Report Author(s): CORTNEY WEINBAUM, JOHN V. PARACHINI, RICHARD S. GIRVEN,
MICHAEL H. DECKER and RICHARD C. BAFFA
RAND Corporation (2018)

Stable URL: http://www.jstor.com/stable/resrep20002.5

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

RAND Corporation is collaborating with JSTOR to digitize, preserve and extend access to this
content.

This content downloaded from

185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
Chapter 3. Unifying Tasking, Collection,
Processing, Exploitation, and Dissemination
(TCPED) Across the U.S. Intelligence Community
T
he IC’s topline budget amount for fiscal year
(FY) 2016, including both base and supplemental
budgets, totaled $70.7 billion.1 According to some
experts, the United States spends more on collec-
tion, analysis, and dissemination of intelligence than the
rest of the world’s intelligence organizations combined.2 In
addition to vast financial resources, the U.S. government
has access to the most-sophisticated intelligence collection
technologies in the world; recruits, trains, and sustains a
highly educated, motivated, and talented workforce; and
currently functions at a higher level of sharing, transpar-
ency, and cooperation than at any previous time in U.S.
history. With that said, a limiting factor in our nation’s
ability to realize the full potential of its combined intel-
ligence strengths and resources resides in the disaggre-
gated and nonfederated tasking, collection, processing,
exploitation, and dissemination (TCPED) architectures
and processes used.3 TCPED is the backbone of the IC, the
framework on which all intelligence rests. It should be the
aggregate of foundational systems and processes that allow
leaders to seek intelligence answers to questions, allocate
resources, direct analysis and production of finished intel-
ligence, and deliver it to the appropriate end users at the
appropriate time.
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
Prioritization of intelligence-gathering resources is
only loosely associated with the National Intelligence
Priorities Framework (NIPF), the Office of the Director of
National Intelligence’s (ODNI’s) “primary mechanism to
establish, disestablish, manage, and communicate national
intelligence priorities.”4 Tasking for collection across the
various intelligence-gathering disciplines is stovepiped
within individual agencies and integrated only loosely
at the national level. Processing and exploitation of col-
lected data remain largely disaggregated, while access
to “raw” collection remains limited at best. In practice,
the intelligence-gathering disciplines, or “INTs,” are not
resourced in direct accordance with the NIPF. The NIPF
serves as a rough set of guidelines to be cited when useful,
but it is only loosely associated with the day-to-day budget
priorities of the IC. The NIPF only marginally impacts the
routine tasking of individual INTs within their silos by
INT functional managers at the agencies who have that
authority—human intelligence (HUMINT) and OSINT
at CIA, signals intelligence (SIGINT) at the National
Security Agency (NSA), geospatial intelligence (GEOINT)
at the National Geospatial-Intelligence Agency (NGA),
and measurement and signature intelligence (MASINT) at
DIA—or by National Intelligence Managers at the national
16
level (who can look into their own regional or functional
silo but have little impact on resources or TCPED decisions
across the entire enterprise). Decentralized acquisition and
disconnected operations of TCPED architectures further
limit processing, exploitation, and dissemination across
the community. For example, sensors flown on unmanned
aerial vehicles acquired and operated by the military
services often do not connect to existing national-level
IC architectures, preventing analysts across the IC from
benefiting from all data sources. In addition, the U.S. space
and ground architecture includes too few nodes where
satellite data can be received, processed, and transmitted to
global users in a timely manner. Meanwhile, disconnects
across classified domains prevent warfighters on ships, in
cockpits, and at forward operating bases from accessing the
most current intelligence in real time.
Discussions of what is required to improve intelligence
often focus on increasing topline resources. IC officials
have historically argued that the IC budget is too small,
considering that the IC cannot or struggles to answer many
policymaker questions on a timely basis.5 Other experts
argue that additional dollars do not necessarily translate
directly into improved intelligence. In a progress review of
the Intelligence Reform and Terrorism Prevention Act of
2004 (IRTPA), Richard Best of the Congressional Research
Service cautioned members of Congress to temper their
expectations: “It should be remembered that intelligence
analysis is an intellectual exercise; it is not possible to
increase budgets by 50 percent and receive 50 percent bet-
ter analysis in the next fiscal year.”6
One of the things required to improve intelligence
is neither growth nor diminution of the IC’s budget but
rather a federating and unifying of the collective and
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
individual components within the TCPED processes of the
17 organizations of the IC,7 as well as an expanded under-
standing of how TCPED applies to every INT in the IC at
both the operational warfighting and strategic decision-
making levels.8 The IC could develop a federated approach
to TCPED that manages centrally but allows for decentral-
ized execution of all resources across the community. Such
an approach would task and integrate across all collection
platforms; make use of all available resources in the com-
munity; and allow streamlined access to raw, processed,
and analyzed data across the entire intelligence enterprise.
Below, we explore how such a process might work, and how
it might address some of the problems with the current
approach to TCPED.
Federated Tasking?
Joint Publication 2.0 describes how TCPED supports, or
should support, the Joint Force Commander’s (JFC’s) abil-
ity to achieve operational success over an adversary:
To prevail, the JFC’s decision and execution cycles
must be consistently faster than the adversary’s and
be based on better information. Being faster and bet-
ter requires having unfettered access to the tasking,
collection, processing, analysis, and dissemination of
information derived from all available sources.9
Similarly, the availability of intelligence information
can be crucial to strategic decisionmaking at the national
and operational levels.
Unfortunately, tasking for either collection or analytic
production across the intelligence enterprise is far from
federated, and commanders and decisionmakers seldom
17
have access to “all available sources.” On any given day, IC
analysts from one organization may be seeking the answers
to an intelligence question that was answered by collection
or analysis conducted in another organization days, weeks,
or even months earlier. This duplication of effort occurs
because responses to ad hoc collection or production tasks
are generally not cataloged or codified inside a single
organization, let alone across the entire IC, and thus are
not discoverable. Improving such discoverability might be
accomplished by improving technological systems or oper-
ational processes, but it might also be improved by recon-
sidering the actual organization and individual missions of
the component agencies of the IC.
While duplication of effort is itself a problem, an
equally serious issue is the possibility that different answers
might be provided to the same question. Although in some
cases multiple analyses of the same data by separate orga-
nizations can provide alternative analyses for consideration
by commanders and policymakers, decisionmakers are
often frustrated by duplicative, repetitive, or contradictory
intelligence analyses that make their consideration and
The same system that
results in duplicative
analyses could also
prevent analysis from
reaching those who need it.
18
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
decisionmaking more time-consuming and sometimes
more difficult.10
The same system that results in duplicative analyses
could also prevent analysis from reaching those who need
it. There are many reasons that access to information might
be blocked: Some information remains stovepiped because
of real or perceived requirements to protect sources and to
ensure continued access to the information, some analytic
products are created in-house for organizational leaders
only and never shared beyond the organization’s walls, and
other information is marginalized or relegated to long-term
internal storage by the originating organization because
it is perceived as not germane to other organizations’
missions.
Theoretically, in a “perfect universe,” all collection and
production tasking could be federated at the national level
and passed to the organization with the most efficient and
effective means of collecting or producing against the task.
The results of all analyses would be made available to all
with a need to know, such that decisionmakers would have
access to “all available sources.”
In the real world, however, where intelligence resources
are limited, the volume and variety of sources of informa-
tion are nearly infinite, and the ability to determine which
organization’s collection methods, assets, sources, or ana-
lysts are best suited to a particular task, federation remains
beyond the IC’s grasp except in limited circumstances
where IC-wide surges or task forces divide the labor across
multiple organizations. Some federation occurs among
functional managers within the INTs, national intelligence
managers within regions and functions, and IC organiza-
tions at the macro level; true federation, however, remains
elusive.
Too Much Collection?
Some experts express concerns that the IC collects too
much information—more than it can process, translate, or
analyze, and certainly more than it can effectively utilize.
Leaving aside recent arguments about excessive intelligence
collection being a threat to civil liberties and privacy, col-
lecting more information than the IC enterprise can rea-
sonably process may be a waste of precious resources and
likely adds to the burden of analysts and analytic tools that
are already overtaxed trying to process and make sense of
the volume of data pouring in daily.11
While it is true that the volume of information now
available in the world exceeds the IC’s capability to collect
and process, it does not necessarily follow that there needs
to be a reduction in collection. If you are looking for a nee-
dle in a haystack, you need to collect the entire stack of hay
to find the needle. It does not matter how many people or
tools you use to sift through a half-stack—if the needle is in
the other half, you will never find it. In the case of the IC,
analysts are looking for tens of thousands of needles every
day in billions of incomplete haystacks across 17 organiza-
tions and within hundreds, potentially thousands, of inter-
nal silos of information.12 Collecting only the information
required is much more difficult than collecting it all and
sifting through it after the fact. The answer to the reputed
problem of “too much collection” resides not in reducing
information intake, but in federating and expanding the
capabilities, tools, and processes used to manage and task
searches within the various “haystacks,” and to exploit and
disseminate the “needles” as they are found.
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
Collecting only the
information required is
much more difficult than
collecting it all and sifting
through it after the fact.
Stovepiped Processing and
Exploitation?
Federating the stovepipes of collection and processing
within the IC is not a new idea. The House Permanent
Select Committee on Intelligence (HPSCI) released a staff
study in April 1996 titled IC21 that spoke directly to this
issue:
The most common criticism of the current collection
management process, and one in which we concur,
is that it is dominated by “stovepipes,” i.e., types of
collection that are managed so as to be largely dis-
tinct from one another. There are several net results.
First, the collection disciplines become competi-
tors for resources driven as much by bureaucratic
imperatives as by a broader national need. Second, it
also becomes much more difficult to make educated
IC-wide decisions about overall collection needs and
the resources required to implement them.13
Some improvements have been made across the IC in
discoverability and access to information since the prom-
ulgation of ICD 501,14 but processing and exploitation of
19
data remain largely stovepiped, both in storage location
and accesses through “stewards” and in the capabilities and
tools used to process the available information.15 What the
HPSCI observed more than two decades ago as a problem
involving increasing collection management and analysis
has expanded into a larger issue of stovepiped processing
and exploitation.
ICD 501 requires, with some exceptions, that IC ele-
ments use “automated means” to make “discoverable” to
authorized IC users all intelligence and intelligence-related
information that they are authorized to “acquire, col-
lect, hold, or obtain,” or analysis that an IC element is
authorized to produce.16 This directive was promulgated
in response to congressional mandates in the IRTPA to
“strengthen the sharing, integration, and management of
information within the Intelligence Community.” However,
realities within the IC suggest that discoverability is far
from perfect, access and processing in real time is all but
nonexistent, and automated means of discovering another
agency’s most closely held or originator-controlled secrets
are still hampered by a system of stewards, gatekeepers,
and internal processing decisions.
Consider SIGINT collected by NSA from a foreign
source and in a foreign language. NSA possesses processing
tools, language analysts, and analytic capabilities to locate,
translate, process, and disseminate in English the infor-
mation that it collects for use by analysts in the rest of the
IC. But despite NSA’s exquisite capabilities, only a fraction
of the SIGINT it collects ever gets processed, translated,
or disseminated.17 While analysts in other agencies might
have language skills that would allow them to help process
and analyze the data NSA collects, most have very limited
access to NSA’s “raw” collection. Increasingly, advanced
20
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
machine reading capabilities, search algorithms, automated
translation, image detection, and the like are making
human first-looks less important, but until we get to a point
where the machine tells us when it has collected something
we are interested in, we should use all the human capability
we have at our disposal.
Similarly, while new tools and applications have
allowed NGA to exploit and disseminate increased
amounts of imagery for use in the IC, the volume of imag-
ery collected each day far exceeds the capacity of NGA
analysts to view, analyze, or comment on all but the most
critical, time-sensitive images. All-source analysts in other
IC agencies continue to have limited access to unprocessed
or not-yet-released imagery collected by NGA, and even
less access to images or data collected elsewhere by U.S.
intelligence, surveillance, and reconnaissance (ISR) assets.
Add to the volume of data collected by national
technical means the even greater volume of data available
through commercial ISR platforms and openly available
on the internet, or crowdsourced and uploaded as needed
by billions of smartphones and other sensors around the
planet—and it becomes even more obvious that analysts, let
alone commanders and decisionmakers, do not have access
to all available sources of information.
Absence of a Central Dissemination
Clearinghouse
The 1996 HPSCI IC21 Staff Study suggested that the lines
between single-source analysis (such as SIGINT and
IMINT) and all-source analysis were beginning to blur
and that there needed to be greater clarity in analytic
roles for each of the INTs and “in relationship to one
another.”18 Today, the problem still exists but might be
better described as a need to refine the roles that individual
analysts must play regardless of where they sit in the IC’s
organizational chart.
Many of the unique skills and missions that were once
the purview of individual agencies have begun to blend
together. NGA is responsible for providing GEOINT, “the
exploitation and analysis of imagery and geospatial infor-
mation that describes, assesses and visually depicts phys-
ical features and geographically referenced activities on
the Earth.”19 NSA provides SIGINT, “foreign intelligence
from communications and information systems,” for use
by decisionmakers across the U.S. government.20 Given this
division of responsibilities, is a geo-rectified digital feed
that includes foreign language narration and launch video
at a foreign missile installation the purview of NGA, NSA,
or both? And if an analyst at CIA or DIA has the tools and
language capability to process and analyze the information
before NSA or NGA can get to it, should that information
not be made available for exploitation and dissemination
sooner rather than later?
ODNI was created to oversee the 17-organization IC
and “improve information sharing, promote a strategic,
unified direction, and ensure integration across the nation’s
IC.”21 Yet, apart from long-term assessments produced
by the NIC, which falls under the purview of ODNI, the
IC lacks truly centralized mechanisms for disseminat-
ing the nation’s most timely and relevant assessments.
Policymakers and warfighters must sift through and sort
hundreds of daily assessments and determine on a contin-
uous, individual basis which assessments to believe and
which to set aside as “alternative.”
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
The IC lacks truly
centralized mechanisms
for disseminating the
nation’s most timely and
relevant assessments.
Unified TCPED: The Future of
Intelligence
As new technologies emerge and improvements are made—
in machine learning and machine translation, artificial
intelligence, big data sorting and processing capabilities,
still image and video facial recognition, change detection
algorithms and other processing, and exploitation and
analytic tools not yet imagined—analysts could become
ever more capable of working with, exploiting, process-
ing, and analyzing even greater volumes of information
and producing and disseminating higher-quality and
more-timely intelligence analyses. In an IC of the future—
fully networked and connected to all available sources
of information—individual analysts and analytic teams
might have the flexibility to use the skills and tools at their
disposal to respond to decisionmakers’ most critical intelli-
gence needs.
In the interconnected, “internet-of-things” 21st cen-
tury, it might be time to question why geographic location
or agency designations should matter at all. In the black-
and-white analog IC of the not-too-distant past, imagery
21
analysts squinted through lenses at backlit photographs
to interpret imagery nuances that would be missed by the
untrained eye of an all-source analyst. Similarly, crypto-
logic language analysts would hit replay repeatedly as they
struggled to differentiate foreign language and nuanced
meaning from background noise in voice recordings.
Today, however, nearly all data are digital; exploitable;
capable of being processed by myriad tools, techniques,
and technologies; and easily shareable with allies and
friends around the world as needed.
While individual IC agencies continue to believe
that what they collect and the sources and methods they
use should be protected from wide dissemination, the IC
should be as agnostic about where data are processed or
exploited as decisionmakers of the future are likely to be
about the individual agency provenance of their intelli-
gence feeds. Decisionmakers require relevant, timely, and
The IC should be as
agnostic about where data
are processed or exploited
as decisionmakers of
the future are likely to
be about the individual
agency provenance of
their intelligence feeds.
22
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
accurate intelligence to make informed decisions, but they
should not have to use their limited time to ponder which
agency seal to revere above others when presented with
alternatives.
Jack Davis, in his occasional papers on the “founder”
of modern U.S. intelligence analysis, Sherman Kent,
suggested that Kent would say that it is the first responsi-
bility of IC analysts to accommodate clients by producing
assessments timed to their decision cycle and focused on
their learning curve.22 Kent would also urge, Davis asserts,
that analysts “allow time for Directorate, Agency, and,
when appropriate, Community coordination” to permit
challenges to and refinement of data and to accommodate
“collective responsibility” in the IC.23 Within these two
thoughts lies the notion that it is the first responsibility of
the IC, writ large, to accommodate clients by producing
collective assessments, where appropriate, timed to their
decision cycles.
The highest order of intelligence produced for U.S.
decisionmakers has been collective in nature. National
Intelligence Estimates, IC memoranda, and Sense of the
Community memoranda all benefit from the collective
assessment of the IC. In recent years, the President’s Daily
Brief (PDB) evolved for a time to include assessments pro-
duced by analysts across the IC, coordinated with multiple
agencies, and edited and polished by a single, national-level
PDB staff. The highest-ranking decisionmakers in the
nation have traditionally been given the same collective
wisdom of the IC in its entirety, even as many below the
level of Department Secretary or Joint Chiefs of Staff have
received daily books filled with individual agency assess-
ments and predictions.
 Imagine an IC of the future collectively working
against a living and ever-evolving set of intelligence
requirements; collecting and providing unclassified and
classified data in a single, unified system accessible to
analysts with appropriate clearance and need-to-know;
and working through intelligence problems with other
U.S. and possibly allied analysts, wherever they are in the
world, to produce intelligence information that is easily
accessible and discoverable by analysts, warfighters, and
decisionmakers around the globe. For such a fantasy to
ever become reality, agency seals and stovepipes would
have to be permanently replaced with national interest and
collective enterprise in a unified IC TCPED construct.
Concluding Thoughts
Unified TCPED would combine centralized tasking across
multiple subordinate collection organizations with col-
lection management systems that would be “aware” of
other complementary collection efforts without revealing
sources. A greater volume of collected data would be
discoverable and available for processing and exploitation
by analysts and warfighters regardless of their geographic
location or agency affiliation, if such affiliations were
even to remain necessary. This might require creation of
an entity at ODNI enabled with “super user” access to all
sources and methods and empowered to direct unification
of TCPED across the IC. Creation of a TCPED ombuds-
man role to adjudicate inevitable conflicts might also be
considered. Dissemination of analysis would be centralized
to provide the collective best efforts of the IC, but would
continue to include alternative analysis or dissenting views
for consideration without prejudice.
Eventually, even the concept of individual INTs, which
currently compete for limited resources within the IC,
might give way to a larger concept of intelligence domi-
nance through unified TCPED, focused more on finding
needles than building and storing haystacks.

23
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms

Chapter Notes
1
Office of the Director of National Intelligence, “U.S. Intelligence Com-
munity Budget,” 2018.
2
Anne Daugherty Miles, Intelligence Community Spending: Trends and
Issues, Washington, D.C.: Congressional Research Service, November 8,
2016. See also Bernd Debusmann, “U.S. Intelligence Spending—Value
for Money?” Reuters, July 16, 2010.
3
The use of the term federated here refers to a number of organizations
being formed into a single centralized unit, within which each organiza-
tion keeps some internal autonomy.
4
Intelligence Community Directive 204, National Intelligence Priorities
Framework, Washington, D.C.: Office of the Director of National Intelli-
gence, January 2, 2015.
5 inating agency, but much information that does not need to be highly
The FY 2018 National Intelligence Program budget request is 7.5 per-
cent higher than the budget requested for FY 2017.
6
Richard A. Best, Intelligence Reform After Five Years: The Role of the
Director of National Intelligence, Washington, D.C.: Congressional
Research Service, June 22, 2010.
7 Barton Gellman, Dafna Linzer, and Carol D. Leonnig, “Surveillance
Executive Order 12333, United States Intelligence Activities, Washing-
ton, D.C.: The White House, December 4, 1981, amended July 30, 2008.
8 U.S. House of Representatives Permanent Select Committee on Intel-
The major intelligence disciplines considered in this chapter are
HUMINT, SIGINT, GEOINT, MASINT, and OSINT. (Other or sub-
INTs include electronic intelligence [ELINT], technical intelligence
[TECHINT], cyber intelligence [CYBINT], financial intelligence
[FININT], and imagery intelligence [IMINT], but these are not specifi-
cally discussed in this chapter.)
9
Joint Publication 2.0, Joint Intelligence, Washington, D.C.: Chairman of
the Joint Chiefs of Staff, October 22, 2013.
10
Concerns over contradictory and duplicative analyses predate the IC
and the National Security Act of 1947 as amended. For CIA’s account of
President Harry S. Truman’s frustrations on these matters in 1946, see
Central Intelligence Agency, “The Beginning of Intelligence Analysis in
CIA,” undated.
11
Gary Sullivan, “Too Much of a Good Thing,” Baltimore Sun, August 27,
2014. See also Alex Young, “Too Much Information,” Harvard Interna-
tional Review, August 20, 2013.
24
This content downloaded from
185.48.148.204 on Wed, 13 Sep 2023 17:32:11 +00:00
All use subject to https://about.jstor.org/terms
12
Hundreds of internal databases and silos exist within the IC at the
national level, thousands likely exist across the military services and
other subcomponents of the Departments of Defense, State, Homeland
Security, Treasury, etc.
13
U.S. House of Representatives Permanent Select Committee on Intel-
ligence, 104th Congress, IC21: The Intelligence Community in the 21st
Century, June 5, 1996.
14
Intelligence Community Directive 501, Discovery and Dissemination
or Retrieval of Information within the Intelligence Community, Wash-
ington, D.C.: Office of the Director of National Intelligence, January 21,
2009.
15
Analysts at IC agencies have more ability than ever to discover infor-
mation related to their work that may be protected by another agency,
and they have some capability to request release of details from the orig-
protected remains inaccessible. Raw SIGINT or unprocessed IMINT, for
example, is mostly not accessible to analysts throughout the community
except through specific requests.
16
Intelligence Community Directive 501, 2009.
17
Net Yields Few Suspects,” Washington Post, February 5, 2006.
18
ligence, 1996.
19
National Geospatial-Intelligence Agency, “About NGA,” undated.
20
National Security Agency, “Frequently Asked Questions: Signals Intel-
ligence (SIGINT),” last modified May 3, 2016.
21
Office of the Director of National Intelligence, “ODNI Fact Sheet,”
October 2011.
22
Jack Davis, Sherman Kent and the Profession of Intelligence Analysis,
Washington, D.C.: Central Intelligence Agency, Sherman Kent Center
for Intelligence Analysis, Occasional Papers, Vol. 1, No. 5, November
2002b.
23
Jack Davis, 2002b.